Как обучить модель на TensorFlow?

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

from tensorflow import keras

# Загрузить данные
(X_train, y_train), (X_test, y_test) = keras.datasets.mnist.load_data()

# Подготовить данные
X_train = X_train.reshape(X_train.shape[0], 28, 28, 1).astype('float32')
X_test = X_test.reshape(X_test.shape[0], 28, 28, 1).astype('float32')
X_train /= 255
X_test /= 255
y_train = keras.utils.to_categorical(y_train, 10)
y_test = keras.utils.to_categorical(y_test, 10)

# Определить модель
model = keras.models.Sequential([
    keras.layers.Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 1)),
    keras.layers.MaxPooling2D(pool_size=(2, 2)),
    keras.layers.Flatten(),
    keras.layers.Dense(10, activation='softmax')
])
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

# Обучить модель
model.fit(X_train, y_train, batch_size=128, epochs=5, validation_data=(X_test, y_test))

# Оценить модель
test_loss, test_acc = model.evaluate(X_test, y_test)
print('Test accuracy:', test_acc)

# Использовать модель
model.save('model.h5')